Eight strains of Microcystis aeruginosa were isolated from freshwater ponds or water reservoirs at various locations in Taiwan during 1989-1993 (Lee et al., 1998; 1999). All strains
each strain were prepared by mixing 200 mg lyophilized cells in 40 ml methanol with a Vortex and centrifuged to collect the supernatant. Cell residues were repeated for another two extractions by 40 ml methanol. All three methanolic extracts were combined and ready for the preparation of various concentration solutions used in the following assays.
2.1. Mouse toxicity assay
From the ethanolic extract of each M. aeruginosa strain, 14.4, 7.2, 3.6, 1.8, 0.9, 0.45 and 0.225 ml were taken out and transferred into separate vials and evaporated to dryness. These dry residues were re-suspended in 3 ml saline solution (0.95%) for intra-peritoneal injection of mice (ICR strain, male, 20±1 g, from Animal Supply of National Taiwan University Hospital, Taipei, Taiwan), 1 ml each and three duplicates for each dose. This makes seven doses equivalent to 400, 200, 100, 50, 25, 12.5 and 6.25 mg dry cells per kg mouse for mouse toxicity assay.
Four hours after each injection, the mortality of each dose was recorded (Lee et al., 1999).
Toxicities of each M. aeruginosa strain, represented by LD50,were then calculated by probit analysis accordingly (Finney, 1963).
Toxin standard MCYST-LR, prepared as in Lee and Chou (2000) was applied as control in the same assay for its LD50 on ICR strain mice. Six doses equivalent to 200, 100, 50, 37.5, 25
Snowville, UT, USA) one day ahead of the toxicity assay following the method in Lee et al.
(1999). Live larvae were suspended in seawater in a concentration of about 250 individuals per ml and transferred to 96-well-microplates, 50 µl per well. Forty-eight milliliters of the ethanolic
extracts from each M. aeruginosa strains were extracted twice with 24 ml n-hexane to remove the oil soluble components and then evaporated to dryness. Dried residues of each sample were re-dissolved in 0.5 ml seawater to give a concentration of 160 mg/ml in terms of lyophilized cell mass. Further dilutions to concentrations of 80, 40, 20, 10 and 5 mg/ml were also prepared for
every strain. Fifty microliters of each sample solutions were added to each well containing 50 µl of larvae suspension, 4 duplicates of each sample. Mortalities of each well in brine shrimp
assay were recorded after 24 hr treatment and toxicities of each M. aeruginosa strains, represented as LC50, were also calculated by probit analysis (Finney, 1963).
Another batch of the same volume of sample extracts were prepared through the same de-fat procedure as described above and then went through a solid phase extraction (SPE) operation to remove any components that may interfere the death of brine shrimp caused by
microcystins. The SPE operation used a column of 0.6 cm i.d. that was packed with 0.1 g silica gel (Baker silica gel 40 µm flash chromatography packing, J. T. Baker, Phillipsburg, NJ, USA).
Columns were preconditioned with 5 ml of ethyl acetate/isopropanol 4/3 (v/v). De-fatted
re-dissolved in 0.5 ml seawater and then diluted to give concentrations of 160, 80, 40, 20, 10 and 5 mg/ml in terms of lyophilized cell mass separately. Brine shrimp larvae toxicity was thus performed with these solutions in wells on microplate as described above. Toxicities of each M.
aeruginosa strains were calculated and represented by LC50. These data were compared with the data obtained from the same assay without SPE operation.
LC50 of toxin standard, MCYST-LR, was also determined as a control in brine shrimp
larvae toxicity assay for methanolic extracts of M. aeruginosa strains. Five concentrations of MCYST-LR, 20, 10, 5, 2 and 1 µg/ml were applied in the same assay.
2.3. Protein Phosphatase 1 inhibition assay
The PP1 inhibition assay used E. coli recombinant protein phosphatase 1 α-isoform of
rabbit muscle (Cat. No. 539493, CalBiochem, San Diego, CA, USA), in a buffer solution containing 50 mM Tris-HCl, pH7.0, 0.1 mM EDTA, 5 mM dithiothreitol, 0.2 mM MnCl2 and BSA 0.2 mg/ml, following the protocol kindly supported by Dr. Chang (Technical Service, CalBiochem). Enzyme activity assay used p-nitrophenyl phosphate (p-NPP) as substrate and detected by the 405 nm absorption of the reaction product p-NP with a SPECTRA Fluor Plus microplate reader (Tecan, AG, Hombrechtikon, Switzerland).
buffer.
Each concentration of sample extracts (Sam.) including the control (Con.) that contained
no Microcystis extract but methanol of the same ratio in buffer occupied a row of 6 wells separately on a plate. Each well in the row was filled with 100 µl buffer and 50 µl sample solutions. Among the 6 wells in each row, three were added individually with 50 µl enzyme solution and 50 µl buffer as an experiment (Exp.) group in triplication, while the other three were added 100 µl buffer separately as the blank (Blk.) group. Sample solutions were allowed to incubate with the enzyme solutions for 10 min before the adding of 50 µl 250 mM p-NPP
substrate solutions to each well to initiate the enzyme reaction. The absorbance at 405 nm was measured for each well at time zero and time 60 min, and the absorbance difference between these two measurements was then recorded. Average of the absorbance difference of the triplicates in each group was applied in the enzyme activity calculation for each sample groups and control. The term ‘percentage activity of the control’ here is used to describe the remaining enzyme activity after the sample inhibition and its calculation as followed:
1. Activity of Control = average absorbance difference of Exp. (in Con.) – average absorbance difference of Blk. (in Con. )
2. Activity of Sample = average absorbance difference in each well of Exp. (in Sam.) –
Microcystin-LR solutions were also prepared in a successive dilution and they were 1280, 640, 320, 160, 80, 40, 20, 10 and 5 pg per well in the same enzyme activity assay. The same percentage activity of control was plotted against the toxin concentrations to reveal the 50%
inhibition (IC50) of the toxin standard to the protein phosphatase activity. The correlation between the toxin concentration and the percentage activity of control remained was analyzed by several regression methods provided by Microsoft Excel®, and the best-fitted regression formula was applied for the calculation of IC50. Protein phosphatase inhibition activities of the methanolic extracts of various Microcystis strains were also judged by their IC50 through the same correlation analysis. Microcystin contents in cell masses of different Microcystis strains thus can be estimated by converting their IC50 in the protein phosphatase inhibition assay to the amount of MCYST-LR equivalent. The conversion used the ratio of IC50 of MCYST-LR and the IC50 of Microcystis extract, which is multiplied by a factor of 1000 to represent a value equivalent to MCYST-LR (in mg) per gram of dried Microcystis cell mass.
An alternative procedure in protein phosphatase inhibition assay was also developed in this study for direct measurement of MCYST-LR equivalent in Microcystis cell mass. The same protein phosphatase assay was done by diluting the cell extract to a concentration equivalent to
compared and their correlation was judged by the linear regression.
2.4. Correlations among mouse toxicity, Artemia toxicity, and PP1 inhibition assays
To determine whether the toxicities of mice or brine shrimps caused by the methanolic extract of various Microcystis strains were due to their microcystin contents, we first converted the LD50 and LC50 in different toxicity assays into toxin contents in these strains and compare the toxin contents obtained by different assays. The conversion of to the toxin contents of MCYST-LR equivalent followed the same method as described in protein phosphatase inhibition assay, using the LD50 and LC50 of toxin standard, MCYST-LR in their respective assays.
2.5. HPLC analysis of different microcystins in Microcystis
An isocratic HPLC with UV absorption at 238 nm was applied for microcystin analysis.
A Luna phenyl-hexyl 5 µ column (4.6 x 250 mm, Phenomenex, Torrance, CA, USA) and a
mobile phase of 0.01 M ammonium acetate/acetonitrile = 75/25, flow rate 1 ml/min were used for a complete separation of different microcystins. From the mentioned methanolic extracts of Microcystis, 2.4 ml were brought to dryness by N2 stream and re-dissolved in 0.3 ml ethyl acetate/isopropanol = 4/3 (v/v). These solutions went through an SPE sample preparation procedure as described in the Artemia toxicity assay. The collected toxin fractions were dried
[D-Asp3]MCYST-FR, and [D-Asp3]MCYST-WR (Fig. 1), previously purified and identified by Lee et al. (1998) were used to identify the different microcystins in Microcystis. Quantitative
determination of microcystins was based on the calibration curve of MCYST-LR within the range of 0.9 µg and 1 ng.
3. Results
3.1. Mouse toxicity assay
Table 1 shows the results of mouse toxicity assay of MCYST-LR and extracts of various
Microcystis strains. Probit analysis revealed the LD50 of MCYST-LR to be 46.9 µg/kg with the 95% confidence interval of 42.4-51.9 µg/kg. Methanolic extracts of various Microcystis strains
also showed different toxicities against mice, and their LD50 were within the range of 9.8-102.2 mg dry cells/kg of mouse, except strains of M.TN-1 and M.KS-1. These two strains were regarded as nontoxic due to the zero death of mice even after intra-peritoneal injections of extract equivalent to 400 mg cell mass, the highest dose in this experiment. Ranking of the toxic strains based on their potency of toxicity against mice was listed as follows, M.TY-1>M.TN-4>M.TN-3>M.TY-2>M.CY-1>M.TN-2.
on Artemia larvae. Probit analysis showed the LC50 of MCYST-LR at 5.68 µg/ml with 95%
confidence interval between 5.17-6.23 µg/ml. Extract of various Microcystis strains also
showed different toxicities from 0.6 mg to 4.5 mg cells equivalent per ml. Although strain M.KS-1 showed some toxicity to brine shrimp larvae at concentration above 12.5 mg cell mass equivalent/ml, it is regarded as non-toxic as strain M.TN-1. Further purification by solid phase extraction on the de-fatted methanolic extracts of Microcystis showed a reduction of the toxicity of the extracts when they were applied for the same Artemia toxicity assay (Table 2). Solid phase extraction was designed to remove the non-polar component and save the microcystins. It is apparent that the methanolic extract of Microcystis contains components other than microcystins that may kill brine shrimp larvae or reinforce the toxicity of microcystins on brine shrimp. Ranking of the toxic strains of Microcystis according to their toxicities against brine shrimp larvae was listed as follows, M.TY-1>M.TN-4>M.TY-2>M.TN-3>M.TN-2>M.CY-1. The ranking may be slightly different without SPE treatment (Table 2). However, if we divided these six strains into 3 groups of the toxic, less toxic and least toxic, two strains in one group, we could find the consistency in the ranking of groups, in both Artemia and mouse toxicity assay (Table 3).
3.3. PP1 inhibition assay
using the natural logarithm function that best fit the curve of the toxin standard (R2=0.97) was then established. From this curve IC50 of MCYST-LR was determined to be 0.13 ng/ml. Similar calculations have been applied for the corresponding curves of various Microcystis extracts and their IC50 were determined as 25.0 ng/ml (M.TY-1), 198.5 ng/ml (M.TY-2), 241.0 ng/ml (M.CY-1), 457.5 ng/ml (M.TN-2), 196.9 ng/ml (M.TN-3), and 129.5 ng/ml (M.TN-4) (Fig. 2).
3.4. Correlations among mouse toxicity, Artemia toxicity, and PP1 inhibition assays
Toxin contents of various Microcystis strains, in a form of MCYST-LR equivalent, those were converted from LD50 and LC50 in different assay methods were listed in Table 3.
Correlation charts showing the toxin contents obtained from different assays were presented in Fig. 3. It is obvious that the toxin content in the methanolic extracts without SPE treatment is much higher from brine shrimp toxicity assay than those from mouse toxicity, and the regression line diverges from the diagonal line a lot (Fig. 3A). However, with the SPE sample preparation the toxin contents in various Microcystis strains obtained from Artemia toxicity assay were close to those from mouse toxicity assay with slight overestimation (Fig. 3B). Toxin contents converted from IC50 of PP1 inhibition assay showed much agreement with those converted from LD50 of mouse toxicity assay (Fig. 3C), but a direct conversion from the percentage of activity of
both mouse and brine shrimp toxicity were related to their protein phosphatase 1 inhibition caused by the microcystins in Microcystis. However some other toxic effects from other components in Microcystis may be exist, especially in the death of brine shrimp larvae.
3.5. HPLC analysis of microcystins in various strains
From HPLC analysis, a calibration line, y = 830.9x + 4608.3 (R2=0.9939, where y: peak area, x: MCYST-LR in ng) of MCYST-LR within the range of 0.9 µg and 1 ng was obtained from
HPLC analysis, and the HPLC chromatograms of microcystins of various Microcystis strains were presented in Fig. 4. There was no identifiable microcystins could be found in the same profile of nontoxic M.TN-1 and M.KS-1, therefore their chromatograms were excluded here.
Nevertheless, those strains showed toxicity to mice and brine shrimp larvae were found to contain various microcystins (Fig. 4). It was found that strain M.TY-1 contained MCYST-LR mainly, which was more than 85% of its microcystin composition. Quantitative analysis from the calibration curve of MCYST-LR (Fig. 4) revealed the content of MCYST-LR in M.TY-1 was 4.75 ppt. Since there existed other minor microcystins, such as [Dha7]MCYST-LR and [D-Asp3]MCYST-LR (unpublished data) that were not clearly distinguishable in this experiment, the actual MCYST-LR equivalent of M.TY-1 considering the conversion from the toxicity of the extract should thus be a little more than 4.75 ppt. Quantitative analysis of MCYST-LR in other
available at this time, it is not possible to calculate the total toxicity in different strains from HPLC analysis.
4. Discussion
There were two major objectives of this research, one is to determine whether the toxicities of Microcystis against mice or brine shrimp larvae are due to the protein phosphatase inhibition activity of microcystins, and another one is to evaluate the sensitivity and accuracy of these two
toxicity assays and protein phosphatase inhibition assay. It has been reported that the LC50 of MCYST-LR on Artemia salina is 5±0.2 µg/ml (48 hr treatment) (Vezie et al., 1996). We also
reported an LC50 of 22.3 µg/ml of MCYST-LR (Lee, et al., 1999). In this experiment we obtained a value of 5.68 µg/ml for the LC50 of MCYST-LR in a 24 hr treatment. It was found
the brine shrimp eggs used for hatching in both experiments were of the same batch and the hatching rate of this durable egg package had decreased from 90% in 1998 to 30% in 2002. It was speculated the older eggs have a poor viability, so do the larvae from aged eggs. Hence a less LC50 of MCYST-LR or in other word, more sensitive while using larvae from aged eggs in the Artemia toxicity assay. The age of the eggs should be controlled while comparing the LC50 of microcystins or extract of Microcystis on brine shrimp larvae.
preparation procedure of SPE for HPLC analysis was very useful for the removal of interfering substances in the methanolic extract of Microcystis. With the same conversion of LD50 to the toxin content of MCYST-LR equivalent in mouse toxicity assay, it was found that the toxin content obtained from mouse assay had a better correlation with the toxin content obtained from Artemia assay with SPE sample preparation than the toxin content obtained from Artemia assay
without SPE sample preparation (Fig. 3a, b).
Similar conversion was also applied for IC50 of different Microcystis extracts in PP1 inhibition assay, and a very good correlation between the toxin contents obtained from mouse toxicity assay and PP1 inhibition was observed. This result indicated that the mouse toxicity caused by microcystins was directly related to their inhibitions on the protein phosphatase 1 without any other toxic effects from microcystins or other unknown components. It has been reported that the death of mouse is due to the massive hepatic haemorrhage as a result of the liver cytoskeleton disruption caused by the protein phosphorylation imbalance from inhibition of serine/threonine protein phosphatase 1 and 2A by accumulated microcystin in liver (Falconer et al., 1981; Eriksson et al., 1990; Honkanen et al., 1990). However, in the Artemia toxicity assay, the death of brine shrimp larvae may be also from protein phosphatase 1 inhibition of microcystins (a slight correlation between the Artemia and mouse toxicity assay of the sample
were observed to exert toxic effect on brine shrimps. It has been reported that long-chain unsaturated fatty acids are Na+/K+-ATPase inhibitors in brine shrimps (Morohashi et al., 1991).
Bury et al. (1998) also reported that long chain fatty acids in Microcystis aeruginosa were the cause of fish death due to their potent inhibitory effects on fish gill Na+/K+-ATPase. We speculate similar components that may kill brine shrimp larvae or intensify the toxicity of microcystins are in our Microcystis strains when brine shrimps were immersed in the solutions of Microcystis extract, and these components can be removed by SPE sample treatment, especially
in the strains of M.TN-4 and M.TN-3 (Table 3). Hence, there would be a better correlation between the Artemia and mouse toxicity, if the Microcystis extracts went through the SPE procedure before adding to the assay media.
Our practice of PP1 inhibition assay has proven itself a fast and reliable method for quantitative analysis of microcystin when comparing the results from PP1 inhibition assay with the results from other methods (Table 3). Toxin contents obtained as MCYST-LR equivalent were quite different from the data obtained by our previous ELISA analysis (Yu et al., 2002). It was realized that strains of Microcystis might not maintain their toxin content in different batches of culture. We found in this experiment that most of the strains produce more toxins than before,
per well. So, an alternate methodology has been designed and tested. Instead of testing the sample in series concentrations to obtain the IC50 for toxin content conversion, we applied the quantity of extract equivalent to 200 ng dry cells per well in this PP1 assay. From its percentage activity of control MCYST-LR equivalent of the sample was converted based on the regression curve showing the correlation of percentage activity of control and the dose of pure MCYST-LR standard (Fig 2). It was found that such a conversion might give accurate data when the sample concentration was adjusted to have a percentage activity of control falling within the range of 40% to 80%.
Although the detection limit of the PP1 assay applied in the study can reach 5 pg/well for
pure MCYST-LR, same power is only equivalent to 25 ppm when 200 ng dry cell/well was applied. The sensitivity is beyond the action limit of 1.0 µg MCLR/g AFA regulated by the
State of Oregon Department of Agriculture (MCLR: microcystin-LR, AFA: Aphanizomenon flos-aquae a cyanobacteria food supplement) (Schaeffer et al., 1999). Increasing cell amount or
the related algal food supplement has been tried in PP1 assay in order to reach the needed 1 ppm toxin level for positive detection. However, components other than MCYST in samples might interfere with the detection or nonspecifically deactivate PP1 activity that gave false positive results. It was observed when extract of 60 µg or higher dry cell equivalent was applied in one
analysis, unpublished data), an amount up to 250 µg dry cell equivalent per well was proven to be
free from the non-MCYST inhibition. The detection sensitivity then can reach to 20 ppb, equal or higher than the sensitivity of ELISA (Yu et al., 2002).
Another phenomenon is also observed while comparing the non-toxic samples with the control, the absorbance of 405 nm from the product of PP1 reaction was slightly enhanced in the tests of nontoxic Microcystis extract or algal food supplement. Similar observation was also reported by Honkanen et al. (1996), in which PP-2A was applied on the detection of okadaic acid contained in mussels. The increase of absorbance was speculated to be from some unknown phosphatase activators indigenous to samples or trace components that gave absorption at 405 nm.
Another phenomenon is also observed while comparing the non-toxic samples with the control, the absorbance of 405 nm from the product of PP1 reaction was slightly enhanced in the tests of nontoxic Microcystis extract or algal food supplement. Similar observation was also reported by Honkanen et al. (1996), in which PP-2A was applied on the detection of okadaic acid contained in mussels. The increase of absorbance was speculated to be from some unknown phosphatase activators indigenous to samples or trace components that gave absorption at 405 nm.